analog_sa said:
No, no, that's not what I meant. Capacitors will not halve the voltage swing per se, but the voltage swing of the amplifier will still be halved when using capacitors between the DAC and the amplifier. This is not because of the capacitors, it's because the DAC only outputs 0V to 5V.
Oh, and I don't like capacitors for sonic reasons 🙂
If the magnitude of the swing is 5v pk to pk on the dac side of the cap , why would it half itself on the output side ?
I think what he means is that each (complementary) DAC output swings 5V pk-pk. If you put that on the primary of a xformer you effectively have 10v pk-pk.
It has nothing to do with cap coupling or not. It only has to do with balanced or single-ended use of the DAC outputs.
Jan Didden
It has nothing to do with cap coupling or not. It only has to do with balanced or single-ended use of the DAC outputs.
Jan Didden
janneman said:
That is one interpretation and it makes no more sense to me than the OP. The point of a balanced setup is to amplify the difference between the two phases. A unity gain balanced input stage will give the same output swing as either one of the two phases.
I don't know how to explain this any better than I already have:
When I say voltage swing, I mean maximum voltage swing. The +/- 25V single ended amplifier from the previous examples have a maximum voltage swing of 50V (from -25V to +25V). By feeding it from a DAC that only outputs 0V to 5V we are effectively halving its maximum voltage swing since it can never go below 0V (it can only swing from 0V to 25V). Adding a capacitor between the DAC and amplifier does not change this. It can still never go below 0V. Hence the maximum voltage swing is still halved.
Note that this applies only when feeding the positive and negative output of the DAC into two separate single ended amplifiers. With a differential amplifier there would be no problem, but I don't want to limit myself to using only differential amplifiers.
Cauhtemoc said:
When I say voltage swing, I mean maximum voltage swing. The +/- 25V single ended amplifier from the previous examples have a maximum voltage swing of 50V (from -25V to +25V). By feeding it from a DAC that only outputs 0V to 5V we are effectively halving its maximum voltage swing since it can never go below 0V (it can only swing from 0V to 25V). Adding a capacitor between the DAC and amplifier does not change this. It can still never go below 0V. Hence the maximum voltage swing is still halved.
Note that this applies only when feeding the positive and negative output of the DAC into two separate single ended amplifiers. With a differential amplifier there would be no problem, but I don't want to limit myself to using only differential amplifiers.
It's life Jim but not as we know it.
The dac does not swing from 0 to 5v it swings around 2.5V. What the cap does is block that 2.5v DC reference and allowing its replacement with one of your choosing most likely 0v. It has been this way for a long time now.
The dac does not swing from 0 to 5v it swings around 2.5V. What the cap does is block that 2.5v DC reference and allowing its replacement with one of your choosing most likely 0v. It has been this way for a long time now.
I'm failing to see how that would work. If the DAC outputs 0V to 5V, how can a capacitor change this to -2.5V to +2.5V?
Cauhtemoc said:
Alas, that is more of a problem for you than it is for the electronics community at large. You need to understand the relationship between the AC and DC components.
Let me give you an example. We will use one of the outputs from your DAC. 0 to 5 volts.
Say we have a +12v and - 12v supply to play with.
We place a 100 uF capacitor between the Dac output and node1
we place a 100k resistor from node1 to the -12v supply.
We now have a voltage on node1 that varies from -14.5 to -10.5
This is because the capacitor has an infinite resistance to DC, but a small finite one at AC.
We can get this result for any reference voltage. Grounding the 100k resistor would give +2.5v to -2.5v.
I recommend that if you can't see how this works, you need to find a basic electronics book, or find a mentor.
Regards;
Doug
DougL said:
I stand corrected then. Although with the resistor connected to -12V the output will vary from -14.5V to -9.5V, not -10.5V. 😉
My original problem still stands however, because I don't want to use capacitors in the signal path.
Cauhtemoc said:
Would you accept an OPA1632 that converts your balanced to single ended and also provides the output buffer/drive function?
You would get +/-12V peak, but I understand you don't really need that +/-25V, that was just an example?
Another option is to connect your DAC outputs via, say, 1k to a resistor of 5k to -12.5V That would give you 0V DC and abouth 2.25V peak at the resistor junction. (Note: the 12.5V and the second resistor can be tweeked to what you have, BUT that 12.5 or whatever needs to be superclean and stable because it gets on your signal).
Jan Didden
Cauhtemoc said:
It's not clear whether you have much choice in this. How stable is your dac's offset? Using an additional PS to compensate the offset will surely involve new capacitors (electrolytics!) in the signal path. And without a servo of sorts what is to guarantee the stability of the offset?
Capacitors may well be the least problematic way of dealing with this. Btw is your power amp input cap coupled?
Regarding the opa1632 solution i feel a bit sceptical. Very few opamps are anywhere close to transparent. What are the chances the 1632 is one of those?
janneman said:
That wouldn't be much different than using an op amp.
janneman said:
This was mentioned earlier. Feeding the DAC with -2.5V and +2.5V (instead of 0V and 5V) accomplishes the same thing but without the added noise of the extra power supply.
analog_sa said:
It's about as stable as the power supply feeding the DAC. I had planned to use a Jung super regulator for this, so it should be quite stable.
analog_sa said:
I want to be able to use this DAC with any amplifier, whether they have a capacitor coupled input or not.
Cauhtemoc said:
Well you would have that extra -2.5V supply, no?
Problem is that is VERY difficult to keep the two supplies equal in voltage, any difference will immediately become offset. On the other hand, the offset of the + and - DAC outputs IS identical as they come from the same reference. No offense meant, but I think you don't quite grasp what you're up against.
Jan Didden
janneman said:
I am aware of this problem. I had not planned to remove the offset entirely, only decrease it somewhat. If I can get the two supplies matched within 100 mV I would be happy. It's a differential DAC so it's still the difference between the two outputs that matter, if both outputs are offset by 100 mV the difference is still the same.
100 mV offset is not ideal, it would limit the maximum gain to 9 and the maximum voltage swing to 45V on the amplifier from the previous examples, but it's a lot better than a gain of 5 and a maximum voltage swing of 25V.
But I'm open to suggestions, as long as they don't involve op amps, capacitors or transformers. 😀
Cauhtemoc said:
Well, obviously you have a very specific aim in mind. Recognizing that most power amps clip hard at 1 or 2V input, I'm at a loss about all this talk of 25V or more peak output level. But I guess you have all info you need to make a decision. You just have to pick the best compromise.
Jan Didden
I know I can hear the groaning in advance...... But have you tried the 5534D opamp as per the AKM4396 eval board? I tried about a dozen of the popular audiophile opamps discussed on this board, but in the end the cheap trusty old 5534D worked the best. There so inexpensive, they are worth a shot.
I had a chance to talk with an engineer at AKM who worked on the Transporter, and after they spent a fortune evaluating opamp choices for the output buffer/filter they came to the same conclusion. It sounded better with this DAC, and not merely measure the best - which it did.
Just another data point to consider.
-David
I had a chance to talk with an engineer at AKM who worked on the Transporter, and after they spent a fortune evaluating opamp choices for the output buffer/filter they came to the same conclusion. It sounded better with this DAC, and not merely measure the best - which it did.
Just another data point to consider.
-David
janneman said:
You have misunderstood me. I never meant for the DAC to have a voltage swing of more than 5V (the entire purpose of this thread was to find a way to have the DAC output -2.5V to +2.5V). The 25V output from a theoretical power amplifier (first mentioned as an example in post #7).
I still want to try to make a differential to differential converter from discrete components (basically two differential to single ended converters, with one of them being inverted). I have tried everything from a long tailed pair to god knows what. The problem seems to be that either they do not swing equally in both directions (positive and negative), or they behave like a comparator and instantly swings from one end to the other.
dw8083 said:
I talked to AKM a while back and they did indeed try a great number of op amps for the AK4396 before settling on the NE5534. Despite of what most people think it's still a decent op amp, and it work very well with the AK4396. I do not want to use an op amp however.
- Status
- Not open for further replies.